The Spread of Obesity in Social Networks

Are you friends members of the Marathon Runner’s of America club or the Bratwurst and Philly Cheesesteak club? If the answer is the later, you are much more likely to be obese than the former.

This is the finding of a 2007 NEJM paper by Christakis and Fowler. Obese individuals are more likely to be friends, relatives, or spouses with other obese people (and vice versa). The authors contend that there are 3 explanations for why this could be the case empirically.

1. Homophily. This means that individuals choose to associate with people who look like them. In this case, social networks would not cause obesity, it is just that obese individuals choose to hang out with other obese individuals.
2. Counfounding factors. Siblings have the same genes. Obesity social norms within a particular geographic area may affect friends and family in a similar manner. These unobserved, confounding variables may also be the true cause of why
3. Induction. Social influence and peer effects may effect the obesity level of each person in a group. The authors hypothesize that this explanation to be the major avenue by which social networks affect obesity.

The paper tracks a database of 12,067 individuals over 32 years. The regressions use a lagged dependent variable to eliminate problems of serial correlation.

Results

Let us define the ‘ego‘ the person as the person whose behavior is being analyze and the ‘alter’ as a person connected to the ego by a social network. When the ego’s alter is a friend and becomes obesity, there is a 57% chance that the ego will become obese. This impact is larger for same sex friendships (71% probability of become obese if the alter becomes obese) than opposite sex friendships (effect not different from zero).

How does the obesity of one’s spouse affect the ego’s obesity? According to the authors, “[a]mong married couples, when an alter became obese, the spouse was 37% more likely (95% CI, 7 to 73) to become obese. Husbands and wives appeared to affect each other similarly (44% and 37%, respectively).”

What explains this phenomenon that the alter’s obesity will affect the ego’s obesity. It is possible that the social network as a whole experiences similar life events which affect obesity. However, even when alter’s live geographically far from the ego–and thus likely have different life experience over time–this does not change the effect the alter’s obesity has on the ego. Christakis and Fowler claim that this supports their perception that social norms heavily influence obesity. Also, the spread of smoking behavior does not affect the spread of obesity. One would guess that social networks would have a similar effect on smoking and obesity. The authors claim that this finding, “…suggests that the psychosocial mechanisms of the spread of obesity may rely less on behavioral imitation than on a change in an ego’s general perception of the social norms regarding the acceptability of obesity.”

• Christakis NA, Fowler JH. 2007. “The spread of obesity in a large social network over 32 years.” New England Journal of Medicine. 357(4): 370-379.